gdb/solib-darwin.c

   1 /* Handle Darwin shared libraries for GDB, the GNU Debugger.
   2
   3    Copyright (C) 2009-2024 Free Software Foundation, Inc.
   4
   5    This file is part of GDB.
   6
   7    This program is free software; you can redistribute it and/or modify
   8    it under the terms of the GNU General Public License as published by
   9    the Free Software Foundation; either version 3 of the License, or
  10    (at your option) any later version.
  11
  12    This program is distributed in the hope that it will be useful,
  13    but WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15    GNU General Public License for more details.
  16
  17    You should have received a copy of the GNU General Public License
  18    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  19
  20 #include "defs.h"
  21
  22 #include "bfd.h"
  23 #include "objfiles.h"
  24 #include "gdbcore.h"
  25 #include "target.h"
  26 #include "inferior.h"
  27 #include "regcache.h"
  28 #include "gdb_bfd.h"
  29
  30 #include "solist.h"
  31 #include "solib-darwin.h"
  32
  33 #include "mach-o.h"
  34 #include "mach-o/external.h"
  35
  36 struct gdb_dyld_image_info
  37 {
  38   /* Base address (which corresponds to the Mach-O header).  */
  39   CORE_ADDR mach_header;
  40   /* Image file path.  */
  41   CORE_ADDR file_path;
  42   /* st.m_time of image file.  */
  43   unsigned long mtime;
  44 };
  45
  46 /* Content of inferior dyld_all_image_infos structure.
  47    See /usr/include/mach-o/dyld_images.h for the documentation.  */
  48 struct gdb_dyld_all_image_infos
  49 {
  50   /* Version (1).  */
  51   unsigned int version;
  52   /* Number of images.  */
  53   unsigned int count;
  54   /* Image description.  */
  55   CORE_ADDR info;
  56   /* Notifier (function called when a library is added or removed).  */
  57   CORE_ADDR notifier;
  58 };
  59
  60 /* Current all_image_infos version.  */
  61 #define DYLD_VERSION_MIN 1
  62 #define DYLD_VERSION_MAX 15
  63
  64 /* Per PSPACE specific data.  */
  65 struct darwin_info
  66 {
  67   /* Address of structure dyld_all_image_infos in inferior.  */
  68   CORE_ADDR all_image_addr = 0;
  69
  70   /* Gdb copy of dyld_all_info_infos.  */
  71   struct gdb_dyld_all_image_infos all_image {};
  72 };
  73
  74 /* Per-program-space data key.  */
  75 static const registry<program_space>::key<darwin_info>
  76   solib_darwin_pspace_data;
  77
  78 /* Get the darwin solib data for PSPACE.  If none is found yet, add it now.  This
  79    function always returns a valid object.  */
  80
  81 static darwin_info *
  82 get_darwin_info (program_space *pspace)
  83 {
  84   darwin_info *info = solib_darwin_pspace_data.get (pspace);
  85   if (info != nullptr)
  86     return info;
  87
  88   return solib_darwin_pspace_data.emplace (pspace);
  89 }
  90
  91 /* Return non-zero if the version in dyld_all_image is known.  */
  92
  93 static int
  94 darwin_dyld_version_ok (const struct darwin_info *info)
  95 {
  96   return info->all_image.version >= DYLD_VERSION_MIN
  97     && info->all_image.version <= DYLD_VERSION_MAX;
  98 }
  99
 100 /* Read dyld_all_image from inferior.  */
 101
 102 static void
 103 darwin_load_image_infos (struct darwin_info *info)
 104 {
 105   gdb_byte buf[24];
 106   bfd_endian byte_order = gdbarch_byte_order (current_inferior ()->arch ());
 107   type *ptr_type
 108     = builtin_type (current_inferior ()->arch ())->builtin_data_ptr;
 109   int len;
 110
 111   /* If the structure address is not known, don't continue.  */
 112   if (info->all_image_addr == 0)
 113     return;
 114
 115   /* The structure has 4 fields: version (4 bytes), count (4 bytes),
 116      info (pointer) and notifier (pointer).  */
 117   len = 4 + 4 + 2 * ptr_type->length ();
 118   gdb_assert (len <= sizeof (buf));
 119   memset (&info->all_image, 0, sizeof (info->all_image));
 120
 121   /* Read structure raw bytes from target.  */
 122   if (target_read_memory (info->all_image_addr, buf, len))
 123     return;
 124
 125   /* Extract the fields.  */
 126   info->all_image.version = extract_unsigned_integer (buf, 4, byte_order);
 127   if (!darwin_dyld_version_ok (info))
 128     return;
 129
 130   info->all_image.count = extract_unsigned_integer (buf + 4, 4, byte_order);
 131   info->all_image.info = extract_typed_address (buf + 8, ptr_type);
 132   info->all_image.notifier = extract_typed_address
 133     (buf + 8 + ptr_type->length (), ptr_type);
 134 }
 135
 136 /* Link map info to include in an allocated so_list entry.  */
 137
 138 struct lm_info_darwin final : public lm_info
 139 {
 140   /* The target location of lm.  */
 141   CORE_ADDR lm_addr = 0;
 142 };
 143
 144 /* Lookup the value for a specific symbol.  */
 145
 146 static CORE_ADDR
 147 lookup_symbol_from_bfd (bfd *abfd, const char *symname)
 148 {
 149   long storage_needed;
 150   asymbol **symbol_table;
 151   unsigned int number_of_symbols;
 152   unsigned int i;
 153   CORE_ADDR symaddr = 0;
 154
 155   storage_needed = bfd_get_symtab_upper_bound (abfd);
 156
 157   if (storage_needed <= 0)
 158     return 0;
 159
 160   symbol_table = (asymbol **) xmalloc (storage_needed);
 161   number_of_symbols = bfd_canonicalize_symtab (abfd, symbol_table);
 162
 163   for (i = 0; i < number_of_symbols; i++)
 164     {
 165       asymbol *sym = symbol_table[i];
 166
 167       if (strcmp (sym->name, symname) == 0
 168           && (sym->section->flags & (SEC_CODE | SEC_DATA)) != 0)
 169         {
 170           /* BFD symbols are section relative.  */
 171           symaddr = sym->value + sym->section->vma;
 172           break;
 173         }
 174     }
 175   xfree (symbol_table);
 176
 177   return symaddr;
 178 }
 179
 180 /* Return program interpreter string.  */
 181
 182 static char *
 183 find_program_interpreter (void)
 184 {
 185   char *buf = NULL;
 186
 187   /* If we have an current exec_bfd, get the interpreter from the load
 188      commands.  */
 189   if (current_program_space->exec_bfd ())
 190     {
 191       bfd_mach_o_load_command *cmd;
 192
 193       if (bfd_mach_o_lookup_command (current_program_space->exec_bfd (),
 194                                      BFD_MACH_O_LC_LOAD_DYLINKER, &cmd) == 1)
 195         return cmd->command.dylinker.name_str;
 196     }
 197
 198   /* If we didn't find it, read from memory.
 199      FIXME: todo.  */
 200   return buf;
 201 }
 202
 203 /*  Not used.  I don't see how the main symbol file can be found: the
 204     interpreter name is needed and it is known from the executable file.
 205     Note that darwin-nat.c implements pid_to_exec_file.  */
 206
 207 static int
 208 open_symbol_file_object (int from_tty)
 209 {
 210   return 0;
 211 }
 212
 213 /* Build a list of currently loaded shared objects.  See solib-svr4.c.  */
 214
 215 static intrusive_list<solib>
 216 darwin_current_sos ()
 217 {
 218   type *ptr_type
 219     = builtin_type (current_inferior ()->arch ())->builtin_data_ptr;
 220   enum bfd_endian byte_order = type_byte_order (ptr_type);
 221   int ptr_len = ptr_type->length ();
 222   unsigned int image_info_size;
 223   darwin_info *info = get_darwin_info (current_program_space);
 224
 225   /* Be sure image infos are loaded.  */
 226   darwin_load_image_infos (info);
 227
 228   if (!darwin_dyld_version_ok (info))
 229     return {};
 230
 231   image_info_size = ptr_len * 3;
 232
 233   intrusive_list<solib> sos;
 234
 235   /* Read infos for each solib.
 236      The first entry was rumored to be the executable itself, but this is not
 237      true when a large number of shared libraries are used (table expanded ?).
 238      We now check all entries, but discard executable images.  */
 239   for (int i = 0; i < info->all_image.count; i++)
 240     {
 241       CORE_ADDR iinfo = info->all_image.info + i * image_info_size;
 242       gdb_byte buf[image_info_size];
 243       CORE_ADDR load_addr;
 244       CORE_ADDR path_addr;
 245       struct mach_o_header_external hdr;
 246       unsigned long hdr_val;
 247
 248       /* Read image info from inferior.  */
 249       if (target_read_memory (iinfo, buf, image_info_size))
 250         break;
 251
 252       load_addr = extract_typed_address (buf, ptr_type);
 253       path_addr = extract_typed_address (buf + ptr_len, ptr_type);
 254
 255       /* Read Mach-O header from memory.  */
 256       if (target_read_memory (load_addr, (gdb_byte *) &hdr, sizeof (hdr) - 4))
 257         break;
 258       /* Discard wrong magic numbers.  Shouldn't happen.  */
 259       hdr_val = extract_unsigned_integer
 260         (hdr.magic, sizeof (hdr.magic), byte_order);
 261       if (hdr_val != BFD_MACH_O_MH_MAGIC && hdr_val != BFD_MACH_O_MH_MAGIC_64)
 262         continue;
 263       /* Discard executable.  Should happen only once.  */
 264       hdr_val = extract_unsigned_integer
 265         (hdr.filetype, sizeof (hdr.filetype), byte_order);
 266       if (hdr_val == BFD_MACH_O_MH_EXECUTE)
 267         continue;
 268
 269       gdb::unique_xmalloc_ptr<char> file_path
 270         = target_read_string (path_addr, SO_NAME_MAX_PATH_SIZE - 1);
 271       if (file_path == nullptr)
 272         break;
 273
 274       /* Create and fill the new struct solib element.  */
 275       solib *newobj = new solib;
 276
 277       auto li = std::make_unique<lm_info_darwin> ();
 278
 279       newobj->so_name = file_path.get ();
 280       newobj->so_original_name = newobj->so_name;
 281       li->lm_addr = load_addr;
 282
 283       newobj->lm_info = std::move (li);
 284       sos.push_back (*newobj);
 285     }
 286
 287   return sos;
 288 }
 289
 290 /* Check LOAD_ADDR points to a Mach-O executable header.  Return LOAD_ADDR
 291    in case of success, 0 in case of failure.  */
 292
 293 static CORE_ADDR
 294 darwin_validate_exec_header (CORE_ADDR load_addr)
 295 {
 296   bfd_endian byte_order = gdbarch_byte_order (current_inferior ()->arch ());
 297   struct mach_o_header_external hdr;
 298   unsigned long hdr_val;
 299
 300   /* Read Mach-O header from memory.  */
 301   if (target_read_memory (load_addr, (gdb_byte *) &hdr, sizeof (hdr) - 4))
 302     return 0;
 303
 304   /* Discard wrong magic numbers.  Shouldn't happen.  */
 305   hdr_val = extract_unsigned_integer
 306     (hdr.magic, sizeof (hdr.magic), byte_order);
 307   if (hdr_val != BFD_MACH_O_MH_MAGIC && hdr_val != BFD_MACH_O_MH_MAGIC_64)
 308     return 0;
 309
 310   /* Check executable.  */
 311   hdr_val = extract_unsigned_integer
 312     (hdr.filetype, sizeof (hdr.filetype), byte_order);
 313   if (hdr_val == BFD_MACH_O_MH_EXECUTE)
 314     return load_addr;
 315
 316   return 0;
 317 }
 318
 319 /* Get the load address of the executable using dyld list of images.
 320    We assume that the dyld info are correct (which is wrong if the target
 321    is stopped at the first instruction).  */
 322
 323 static CORE_ADDR
 324 darwin_read_exec_load_addr_from_dyld (struct darwin_info *info)
 325 {
 326   type *ptr_type
 327     = builtin_type (current_inferior ()->arch ())->builtin_data_ptr;
 328   int ptr_len = ptr_type->length ();
 329   unsigned int image_info_size = ptr_len * 3;
 330   int i;
 331
 332   /* Read infos for each solib.  One of them should be the executable.  */
 333   for (i = 0; i < info->all_image.count; i++)
 334     {
 335       CORE_ADDR iinfo = info->all_image.info + i * image_info_size;
 336       gdb_byte buf[image_info_size];
 337       CORE_ADDR load_addr;
 338
 339       /* Read image info from inferior.  */
 340       if (target_read_memory (iinfo, buf, image_info_size))
 341         break;
 342
 343       load_addr = extract_typed_address (buf, ptr_type);
 344       if (darwin_validate_exec_header (load_addr) == load_addr)
 345         return load_addr;
 346     }
 347
 348   return 0;
 349 }
 350
 351 /* Get the load address of the executable when the PC is at the dyld
 352    entry point using parameter passed by the kernel (at SP). */
 353
 354 static CORE_ADDR
 355 darwin_read_exec_load_addr_at_init (struct darwin_info *info)
 356 {
 357   gdbarch *gdbarch = current_inferior ()->arch ();
 358   enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
 359   int addr_size = gdbarch_addr_bit (gdbarch) / 8;
 360   ULONGEST load_ptr_addr;
 361   ULONGEST load_addr;
 362   gdb_byte buf[8];
 363
 364   /* Get SP.  */
 365   if (regcache_cooked_read_unsigned (get_thread_regcache (inferior_thread ()),
 366                                      gdbarch_sp_regnum (gdbarch),
 367                                      &load_ptr_addr) != REG_VALID)
 368     return 0;
 369
 370   /* Read value at SP (image load address).  */
 371   if (target_read_memory (load_ptr_addr, buf, addr_size))
 372     return 0;
 373
 374   load_addr = extract_unsigned_integer (buf, addr_size, byte_order);
 375
 376   return darwin_validate_exec_header (load_addr);
 377 }
 378
 379 /* Return 1 if PC lies in the dynamic symbol resolution code of the
 380    run time loader.  */
 381
 382 static int
 383 darwin_in_dynsym_resolve_code (CORE_ADDR pc)
 384 {
 385   return 0;
 386 }
 387
 388 /* A wrapper for bfd_mach_o_fat_extract that handles reference
 389    counting properly.  This will either return NULL, or return a new
 390    reference to a BFD.  */
 391
 392 static gdb_bfd_ref_ptr
 393 gdb_bfd_mach_o_fat_extract (bfd *abfd, bfd_format format,
 394                             const bfd_arch_info_type *arch)
 395 {
 396   bfd *result = bfd_mach_o_fat_extract (abfd, format, arch);
 397
 398   if (result == NULL)
 399     return NULL;
 400
 401   if (result == abfd)
 402     gdb_bfd_ref (result);
 403   else
 404     gdb_bfd_mark_parent (result, abfd);
 405
 406   return gdb_bfd_ref_ptr (result);
 407 }
 408
 409 /* Return the BFD for the program interpreter.  */
 410
 411 static gdb_bfd_ref_ptr
 412 darwin_get_dyld_bfd ()
 413 {
 414   char *interp_name;
 415
 416   /* This method doesn't work with an attached process.  */
 417   if (current_inferior ()->attach_flag)
 418     return NULL;
 419
 420   /* Find the program interpreter.  */
 421   interp_name = find_program_interpreter ();
 422   if (!interp_name)
 423     return NULL;
 424
 425   /* Create a bfd for the interpreter.  */
 426   gdb_bfd_ref_ptr dyld_bfd (gdb_bfd_open (interp_name, gnutarget));
 427   if (dyld_bfd != NULL)
 428     {
 429       gdb_bfd_ref_ptr sub
 430         (gdb_bfd_mach_o_fat_extract
 431            (dyld_bfd.get (), bfd_object,
 432             gdbarch_bfd_arch_info (current_inferior ()->arch ())));
 433       dyld_bfd = sub;
 434     }
 435   return dyld_bfd;
 436 }
 437
 438 /* Extract dyld_all_image_addr when the process was just created, assuming the
 439    current PC is at the entry of the dynamic linker.  */
 440
 441 static void
 442 darwin_solib_get_all_image_info_addr_at_init (struct darwin_info *info)
 443 {
 444   CORE_ADDR load_addr = 0;
 445   gdb_bfd_ref_ptr dyld_bfd = darwin_get_dyld_bfd ();
 446
 447   if (dyld_bfd == NULL)
 448     return;
 449
 450   /* We find the dynamic linker's base address by examining
 451      the current pc (which should point at the entry point for the
 452      dynamic linker) and subtracting the offset of the entry point.  */
 453   load_addr = (regcache_read_pc (get_thread_regcache (inferior_thread ()))
 454                - bfd_get_start_address (dyld_bfd.get ()));
 455
 456   /* Now try to set a breakpoint in the dynamic linker.  */
 457   info->all_image_addr =
 458     lookup_symbol_from_bfd (dyld_bfd.get (), "_dyld_all_image_infos");
 459
 460   if (info->all_image_addr == 0)
 461     return;
 462
 463   info->all_image_addr += load_addr;
 464 }
 465
 466 /* Extract dyld_all_image_addr reading it from
 467    TARGET_OBJECT_DARWIN_DYLD_INFO.  */
 468
 469 static void
 470 darwin_solib_read_all_image_info_addr (struct darwin_info *info)
 471 {
 472   gdb_byte buf[8];
 473   LONGEST len;
 474   type *ptr_type
 475     = builtin_type (current_inferior ()->arch ())->builtin_data_ptr;
 476
 477   /* Sanity check.  */
 478   if (ptr_type->length () > sizeof (buf))
 479     return;
 480
 481   len = target_read (current_inferior ()->top_target (),
 482                      TARGET_OBJECT_DARWIN_DYLD_INFO,
 483                      NULL, buf, 0, ptr_type->length ());
 484   if (len <= 0)
 485     return;
 486
 487   /* The use of BIG endian is intended, as BUF is a raw stream of bytes.  This
 488       makes the support of remote protocol easier.  */
 489   info->all_image_addr = extract_unsigned_integer (buf, len, BFD_ENDIAN_BIG);
 490 }
 491
 492 /* Shared library startup support.  See documentation in solib-svr4.c.  */
 493
 494 static void
 495 darwin_solib_create_inferior_hook (int from_tty)
 496 {
 497   /* Everything below only makes sense if we have a running inferior.  */
 498   if (!target_has_execution ())
 499     return;
 500
 501   darwin_info *info = get_darwin_info (current_program_space);
 502   CORE_ADDR load_addr;
 503
 504   info->all_image_addr = 0;
 505
 506   darwin_solib_read_all_image_info_addr (info);
 507
 508   if (info->all_image_addr == 0)
 509     darwin_solib_get_all_image_info_addr_at_init (info);
 510
 511   if (info->all_image_addr == 0)
 512     return;
 513
 514   darwin_load_image_infos (info);
 515
 516   if (!darwin_dyld_version_ok (info))
 517     {
 518       warning (_("unhandled dyld version (%d)"), info->all_image.version);
 519       return;
 520     }
 521
 522   if (info->all_image.count != 0)
 523     {
 524       /* Possible relocate the main executable (PIE).  */
 525       load_addr = darwin_read_exec_load_addr_from_dyld (info);
 526     }
 527   else
 528     {
 529       /* Possible issue:
 530          Do not break on the notifier if dyld is not initialized (deduced from
 531          count == 0).  In that case, dyld hasn't relocated itself and the
 532          notifier may point to a wrong address.  */
 533
 534       load_addr = darwin_read_exec_load_addr_at_init (info);
 535     }
 536
 537   if (load_addr != 0 && current_program_space->symfile_object_file != NULL)
 538     {
 539       CORE_ADDR vmaddr;
 540
 541       /* Find the base address of the executable.  */
 542       vmaddr = bfd_mach_o_get_base_address (current_program_space->exec_bfd ());
 543
 544       /* Relocate.  */
 545       if (vmaddr != load_addr)
 546         objfile_rebase (current_program_space->symfile_object_file,
 547                         load_addr - vmaddr);
 548     }
 549
 550   /* Set solib notifier (to reload list of shared libraries).  */
 551   CORE_ADDR notifier = info->all_image.notifier;
 552
 553   if (info->all_image.count == 0)
 554     {
 555       /* Dyld hasn't yet relocated itself, so the notifier address may
 556          be incorrect (as it has to be relocated).  */
 557       CORE_ADDR start
 558         = bfd_get_start_address (current_program_space->exec_bfd ());
 559       if (start == 0)
 560         notifier = 0;
 561       else
 562         {
 563           gdb_bfd_ref_ptr dyld_bfd = darwin_get_dyld_bfd ();
 564           if (dyld_bfd != NULL)
 565             {
 566               CORE_ADDR dyld_bfd_start_address;
 567               CORE_ADDR dyld_relocated_base_address;
 568               CORE_ADDR pc;
 569
 570               dyld_bfd_start_address = bfd_get_start_address (dyld_bfd.get());
 571
 572               /* We find the dynamic linker's base address by examining
 573                  the current pc (which should point at the entry point
 574                  for the dynamic linker) and subtracting the offset of
 575                  the entry point.  */
 576
 577               pc = regcache_read_pc (get_thread_regcache (inferior_thread ()));
 578               dyld_relocated_base_address = pc - dyld_bfd_start_address;
 579
 580               /* We get the proper notifier relocated address by
 581                  adding the dyld relocated base address to the current
 582                  notifier offset value.  */
 583
 584               notifier += dyld_relocated_base_address;
 585             }
 586         }
 587     }
 588
 589   /* Add the breakpoint which is hit by dyld when the list of solib is
 590      modified.  */
 591   if (notifier != 0)
 592     create_solib_event_breakpoint (current_inferior ()->arch (), notifier);
 593 }
 594
 595 static void
 596 darwin_clear_solib (program_space *pspace)
 597 {
 598   darwin_info *info = get_darwin_info (pspace);
 599
 600   info->all_image_addr = 0;
 601   info->all_image.version = 0;
 602 }
 603
 604 /* The section table is built from bfd sections using bfd VMAs.
 605    Relocate these VMAs according to solib info.  */
 606
 607 static void
 608 darwin_relocate_section_addresses (solib &so, target_section *sec)
 609 {
 610   auto *li = gdb::checked_static_cast<lm_info_darwin *> (so.lm_info.get ());
 611
 612   sec->addr += li->lm_addr;
 613   sec->endaddr += li->lm_addr;
 614
 615   /* Best effort to set addr_high/addr_low.  This is used only by
 616      'info sharedlibary'.  */
 617   if (so.addr_high == 0)
 618     {
 619       so.addr_low = sec->addr;
 620       so.addr_high = sec->endaddr;
 621     }
 622   if (sec->endaddr > so.addr_high)
 623     so.addr_high = sec->endaddr;
 624   if (sec->addr < so.addr_low)
 625     so.addr_low = sec->addr;
 626 }
 627 \f
 628 static gdb_bfd_ref_ptr
 629 darwin_bfd_open (const char *pathname)
 630 {
 631   int found_file;
 632
 633   /* Search for shared library file.  */
 634   gdb::unique_xmalloc_ptr<char> found_pathname
 635     = solib_find (pathname, &found_file);
 636   if (found_pathname == NULL)
 637     perror_with_name (pathname);
 638
 639   /* Open bfd for shared library.  */
 640   gdb_bfd_ref_ptr abfd (solib_bfd_fopen (found_pathname.get (), found_file));
 641
 642   gdb_bfd_ref_ptr res
 643     (gdb_bfd_mach_o_fat_extract
 644        (abfd.get (), bfd_object,
 645         gdbarch_bfd_arch_info (current_inferior ()->arch ())));
 646   if (res == NULL)
 647     error (_("`%s': not a shared-library: %s"),
 648            bfd_get_filename (abfd.get ()), bfd_errmsg (bfd_get_error ()));
 649
 650   /* The current filename for fat-binary BFDs is a name generated
 651      by BFD, usually a string containing the name of the architecture.
 652      Reset its value to the actual filename.  */
 653   bfd_set_filename (res.get (), pathname);
 654
 655   return res;
 656 }
 657
 658 const solib_ops darwin_so_ops =
 659 {
 660   darwin_relocate_section_addresses,
 661   nullptr,
 662   darwin_clear_solib,
 663   darwin_solib_create_inferior_hook,
 664   darwin_current_sos,
 665   open_symbol_file_object,
 666   darwin_in_dynsym_resolve_code,
 667   darwin_bfd_open,
 668 };